CITS4009 Computational Data Analysis
Exploratory Data Analysis
Abdul Momen USMANI(23685219)
Semester 2, 2022
Cleaning the Dataset
AccidentDS <- read.csv(file = '/Users/hamzausmani/CITS4009 Project 1 First Sem/us_data_2000.csv')
str(AccidentDS)## 'data.frame': 2000 obs. of 57 variables:
## $ MINE_ID : int 100003 100003 100008 100011 100011 100011 100011 100011 100016 100021 ...
## $ CONTROLLER_ID : chr "41044" "41044" "M31753" "M11763" ...
## $ CONTROLLER_NAME : chr "Lhoist Group" "Lhoist Group" "Alan B Cheney" "Imerys S A" ...
## $ OPERATOR_ID : chr "L13586" "L13586" "L31753" "L17074" ...
## $ OPERATOR_NAME : chr "Lhoist North America " "Lhoist North America " "Cheney Lime & Cement Company" "Imerys Pigments LLC" ...
## $ CONTRACTOR_ID : chr "" "" "" "" ...
## $ DOCUMENT_NO : num 2.2e+11 2.2e+11 2.2e+11 2.2e+11 2.2e+11 ...
## $ SUBUNIT_CD : int 3 30 30 30 30 30 30 30 3 3 ...
## $ SUBUNIT : chr "STRIP, QUARY, OPEN PIT" "MILL OPERATION/PREPARATION PLANT" "MILL OPERATION/PREPARATION PLANT" "MILL OPERATION/PREPARATION PLANT" ...
## $ ACCIDENT_DT : chr "14/03/2012" "8/01/2007" "4/07/2009" "26/05/2000" ...
## $ CAL_YR : int 2012 2007 2009 2000 2005 2006 2008 2012 2000 2006 ...
## $ CAL_QTR : int 1 1 3 2 1 1 4 2 3 1 ...
## $ FISCAL_YR : int 2012 2007 2009 2000 2005 2006 2009 2012 2000 2006 ...
## $ FISCAL_QTR : int 2 2 4 3 2 2 1 3 4 2 ...
## $ ACCIDENT_TIME : int 945 1105 1000 1100 1430 1130 430 930 730 230 ...
## $ DEGREE_INJURY_CD : chr "5" "6" "3" "5" ...
## $ DEGREE_INJURY : chr "DAYS RESTRICTED ACTIVITY ONLY" "NO DYS AWY FRM WRK,NO RSTR ACT" "DAYS AWAY FROM WORK ONLY" "DAYS RESTRICTED ACTIVITY ONLY" ...
## $ FIPS_STATE_CD : int 1 1 1 1 1 1 1 1 1 1 ...
## $ UG_LOCATION_CD : chr "?" "?" "?" "?" ...
## $ UG_LOCATION : chr "NO VALUE FOUND" "NO VALUE FOUND" "NO VALUE FOUND" "NO VALUE FOUND" ...
## $ UG_MINING_METHOD_CD: chr "?" "?" "?" "?" ...
## $ UG_MINING_METHOD : chr "NO VALUE FOUND" "NO VALUE FOUND" "NO VALUE FOUND" "NO VALUE FOUND" ...
## $ MINING_EQUIP_CD : chr "24" "28" "?" "?" ...
## $ MINING_EQUIP : chr "Front-end loader, Tractor-shovel, Payloader, Highlift, Skip loader" "Hand tools (not powered)" "NO VALUE FOUND" "NO VALUE FOUND" ...
## $ EQUIP_MFR_CD : chr "119" "121" "?" "?" ...
## $ EQUIP_MFR_NAME : chr "Not on this list" "Not Reported" "NO VALUE FOUND" "NO VALUE FOUND" ...
## $ EQUIP_MODEL_NO : chr "22321" "" "" "?" ...
## $ SHIFT_BEGIN_TIME : int 600 700 600 700 700 700 2300 700 700 1800 ...
## $ CLASSIFICATION_CD : chr "12" "10" "18" "9" ...
## $ CLASSIFICATION : chr "POWERED HAULAGE" "HANDTOOLS (NONPOWERED)" "SLIP OR FALL OF PERSON" "HANDLING OF MATERIALS" ...
## $ ACCIDENT_TYPE_CD : chr "21" "8" "30" "27" ...
## $ ACCIDENT_TYPE : chr "CGHT I, U, B, MVNG & STTN OBJS" "STRUCK BY, NEC" "OVER-EXERTION, NEC" "OVER-EXERTION IN LIFTING OBJS" ...
## $ NO_INJURIES : int 1 1 1 1 1 1 1 1 1 1 ...
## $ TOT_EXPER : num 4.35 0.02 10 NA 0.87 ...
## $ MINE_EXPER : num 4.35 0.02 2.15 0.23 0.87 ...
## $ JOB_EXPER : num 0.67 0.02 2.15 0.23 0.38 ...
## $ OCCUPATION_CD : chr "374" "374" "374" "374" ...
## $ OCCUPATION : chr "Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator" "Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator" "Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator" "Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator" ...
## $ ACTIVITY_CD : chr "28" "30" "13" "28" ...
## $ ACTIVITY : chr "HANDLING SUPPLIES/MATERIALS" "HAND TOOLS (NOT POWERED)" "CLIMB SCAFFOLDS/LADDERS/PLATFORMS" "HANDLING SUPPLIES/MATERIALS" ...
## $ INJURY_SOURCE_CD : chr "76" "46" "117" "4" ...
## $ INJURY_SOURCE : chr "SURFACE MINING MACHINES" "AXE,HAMMER,SLEDGE" "GROUND" "BAGS" ...
## $ NATURE_INJURY_CD : chr "160" "180" "330" "330" ...
## $ NATURE_INJURY : chr "CONTUSN,BRUISE,INTAC SKIN" "CUT,LACER,PUNCT-OPN WOUND" "SPRAIN,STRAIN RUPT DISC" "SPRAIN,STRAIN RUPT DISC" ...
## $ INJ_BODY_PART_CD : chr "700" "100" "520" "420" ...
## $ INJ_BODY_PART : chr "MULTIPLE PARTS (MORE THAN ONE MAJOR)" "HEAD,NEC" "ANKLE" "BACK (MUSCLES/SPINE/S-CORD/TAILBONE)" ...
## $ SCHEDULE_CHARGE : int 0 0 0 NA 0 0 0 0 NA 0 ...
## $ DAYS_RESTRICT : int 8 0 0 5 5 3 0 21 10 19 ...
## $ DAYS_LOST : int 0 0 9 NA 0 0 0 0 NA 13 ...
## $ TRANS_TERM : chr "N" "N" "N" "N" ...
## $ RETURN_TO_WORK_DT : chr "03/26/2012" "1/09/2007" "07/14/2009" "6/01/2000" ...
## $ IMMED_NOTIFY_CD : chr "? " "? " "? " "13" ...
## $ IMMED_NOTIFY : chr "NO VALUE FOUND" "NO VALUE FOUND" "NO VALUE FOUND" "NOT MARKED" ...
## $ INVEST_BEGIN_DT : chr "" "" "" "" ...
## $ NARRATIVE : chr "Employee was cleaning up at the Primary Crusher with the Dingo skid steer. The employee slipped and fell while "| __truncated__ "Handle of sledgehammer broke and head of hammer hit employee in the forehead." "EMPLOYEE WAS CLIMBING DOWN A LADDER AND WHEN HE STEPPED TO THE GROUND HE SLIPPED AND SPRAINED HIS LEFT ANKLE." "HE PULLED A BACK MUSCLE WHILE STACKING BAGS OF MATERIAL." ...
## $ CLOSED_DOC_NO : num NA NA 3.2e+11 3.2e+11 NA ...
## $ COAL_METAL_IND : chr "M" "M" "M" "M" ...#Identifying the number of nullpoints within the variables and the dimension of the dataset
apply(is.na(AccidentDS), 2, sum)## MINE_ID CONTROLLER_ID CONTROLLER_NAME OPERATOR_ID
## 0 0 0 0
## OPERATOR_NAME CONTRACTOR_ID DOCUMENT_NO SUBUNIT_CD
## 0 0 0 0
## SUBUNIT ACCIDENT_DT CAL_YR CAL_QTR
## 0 0 0 0
## FISCAL_YR FISCAL_QTR ACCIDENT_TIME DEGREE_INJURY_CD
## 0 0 0 0
## DEGREE_INJURY FIPS_STATE_CD UG_LOCATION_CD UG_LOCATION
## 0 0 0 0
## UG_MINING_METHOD_CD UG_MINING_METHOD MINING_EQUIP_CD MINING_EQUIP
## 0 0 0 0
## EQUIP_MFR_CD EQUIP_MFR_NAME EQUIP_MODEL_NO SHIFT_BEGIN_TIME
## 0 0 1 9
## CLASSIFICATION_CD CLASSIFICATION ACCIDENT_TYPE_CD ACCIDENT_TYPE
## 0 0 0 0
## NO_INJURIES TOT_EXPER MINE_EXPER JOB_EXPER
## 0 360 333 328
## OCCUPATION_CD OCCUPATION ACTIVITY_CD ACTIVITY
## 0 0 0 0
## INJURY_SOURCE_CD INJURY_SOURCE NATURE_INJURY_CD NATURE_INJURY
## 0 0 0 0
## INJ_BODY_PART_CD INJ_BODY_PART SCHEDULE_CHARGE DAYS_RESTRICT
## 0 0 673 543
## DAYS_LOST TRANS_TERM RETURN_TO_WORK_DT IMMED_NOTIFY_CD
## 422 0 0 0
## IMMED_NOTIFY INVEST_BEGIN_DT NARRATIVE CLOSED_DOC_NO
## 0 0 0 1138
## COAL_METAL_IND
## 0dimension <- c(ncol(AccidentDS),nrow(AccidentDS))
dimension## [1] 57 2000There are 57 columns and 2000 rows
There are many unnecessary columns which should be removed in order to prevent hampering our visualization
RefinedColNames <- colnames(AccidentDS)Before removing unnecessary columns there are many points to be noted There are many columns which have an equivalent number of same columns such as: CONTROLLER_ID is equivalent to CONTROLLER_NAME OPERATOR_ID is equivalent to OPERATOR_NAME SUBUNIT_CS is equivalent to SUBUNIT The year subset of ACCIDENT_DT is equivalent to CAL_YR DEGREE_INJURY_CD is equivalent to DEGREE_INJURY UG_LOCATION_CD is equivalent to UG_LOCATION UG_MINING_CD is equivalent to UG_MINING_METHOD MINING_EQUIP_CD is equivalent to MINING_EQUIP EQUIP_MFR_CD is equivalent to EQUIP_MFR_NAME CLASSIFICATION_CD is equivalent to CLASSIFICATION ACCIDENT_TYPE_CD is equivalent to ACCIDENT_TYPE OCCUPATION_CD is equivalent to OCCUPATION ACTIVITY_CD is equivalent to ACTIVITY INJURY_SOURCE_CD is equivalent to INJURY_SOURCE NATURE_INJURY_CD is equivalent to NATURE_INJURY INJ_BODY_PART_CD is equivalent to INJ_BODY_PART IMMED_NOTIFY_CD is equivalent to IMMED_NOTIFY
Now we are checking the columns that have missing values
DropColumns<- AccidentDS
drop <- c("MINE_ID","CONTROLLER_ID","OPERATOR_ID","DOCUMENT_NO","SUBUNIT_CD","ACCIDENT_DT","DEGREE_INJURY_CD","UG_LOCATION_CD","UG_MINING_METHOD_CD","MINING_EQUIP_CD","EQUIP_MFR_CD","CLASSIFICATION_CD","ACCIDENT_TYPE_CD","ACCIDENT_TYPE_CD","OCCUPATION_CD","ACTIVITY_CD","INJURY_SOURCE_CD","NATURE_INJURY_CD","INJ_BODY_PART_CD","IMMED_NOTIFY_CD","NARRATIVE","CLOSED_DOC_NO")
df = DropColumns[,!(names(DropColumns) %in% drop)]
head(df)## CONTROLLER_NAME OPERATOR_NAME CONTRACTOR_ID
## 1 Lhoist Group Lhoist North America
## 2 Lhoist Group Lhoist North America
## 3 Alan B Cheney Cheney Lime & Cement Company
## 4 Imerys S A Imerys Pigments LLC
## 5 Imerys S A Imerys Pigments LLC
## 6 Imerys S A Imerys Pigments LLC
## SUBUNIT CAL_YR CAL_QTR FISCAL_YR FISCAL_QTR
## 1 STRIP, QUARY, OPEN PIT 2012 1 2012 2
## 2 MILL OPERATION/PREPARATION PLANT 2007 1 2007 2
## 3 MILL OPERATION/PREPARATION PLANT 2009 3 2009 4
## 4 MILL OPERATION/PREPARATION PLANT 2000 2 2000 3
## 5 MILL OPERATION/PREPARATION PLANT 2005 1 2005 2
## 6 MILL OPERATION/PREPARATION PLANT 2006 1 2006 2
## ACCIDENT_TIME DEGREE_INJURY FIPS_STATE_CD UG_LOCATION
## 1 945 DAYS RESTRICTED ACTIVITY ONLY 1 NO VALUE FOUND
## 2 1105 NO DYS AWY FRM WRK,NO RSTR ACT 1 NO VALUE FOUND
## 3 1000 DAYS AWAY FROM WORK ONLY 1 NO VALUE FOUND
## 4 1100 DAYS RESTRICTED ACTIVITY ONLY 1 NO VALUE FOUND
## 5 1430 DAYS RESTRICTED ACTIVITY ONLY 1 NO VALUE FOUND
## 6 1130 DAYS RESTRICTED ACTIVITY ONLY 1 NO VALUE FOUND
## UG_MINING_METHOD
## 1 NO VALUE FOUND
## 2 NO VALUE FOUND
## 3 NO VALUE FOUND
## 4 NO VALUE FOUND
## 5 NO VALUE FOUND
## 6 NO VALUE FOUND
## MINING_EQUIP
## 1 Front-end loader, Tractor-shovel, Payloader, Highlift, Skip loader
## 2 Hand tools (not powered)
## 3 NO VALUE FOUND
## 4 NO VALUE FOUND
## 5 NO VALUE FOUND
## 6 NO VALUE FOUND
## EQUIP_MFR_NAME EQUIP_MODEL_NO SHIFT_BEGIN_TIME CLASSIFICATION
## 1 Not on this list 22321 600 POWERED HAULAGE
## 2 Not Reported 700 HANDTOOLS (NONPOWERED)
## 3 NO VALUE FOUND 600 SLIP OR FALL OF PERSON
## 4 NO VALUE FOUND ? 700 HANDLING OF MATERIALS
## 5 NO VALUE FOUND 700 HANDLING OF MATERIALS
## 6 NO VALUE FOUND 700 OTHER
## ACCIDENT_TYPE NO_INJURIES TOT_EXPER MINE_EXPER JOB_EXPER
## 1 CGHT I, U, B, MVNG & STTN OBJS 1 4.35 4.35 0.67
## 2 STRUCK BY, NEC 1 0.02 0.02 0.02
## 3 OVER-EXERTION, NEC 1 10.00 2.15 2.15
## 4 OVER-EXERTION IN LIFTING OBJS 1 NA 0.23 0.23
## 5 ABSRTN RAD CAUST TXC & NOX SBS 1 0.87 0.87 0.38
## 6 BODILY REACTION, NEC 1 5.62 5.62 5.62
## OCCUPATION
## 1 Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator
## 2 Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator
## 3 Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator
## 4 Warehouseman, Bagger, Palletizer/Stacker, Store keeper, Packager, Fabricator, Cleaning plant operator
## 5 Front-end loader, Scraper-loader operator, Pan operator, Payloader, Scraper rig operator
## 6 Maintenance man, Mechanic, Repair/Serviceman, Boilermaker, Fueler, Tire tech, Field service tech
## ACTIVITY INJURY_SOURCE
## 1 HANDLING SUPPLIES/MATERIALS SURFACE MINING MACHINES
## 2 HAND TOOLS (NOT POWERED) AXE,HAMMER,SLEDGE
## 3 CLIMB SCAFFOLDS/LADDERS/PLATFORMS GROUND
## 4 HANDLING SUPPLIES/MATERIALS BAGS
## 5 WORKING WITH CHEMICALS ACIDS,ALKALI,WET CEMENT
## 6 MACHINE MAINTENANCE/REPAIR BODILY MOTION
## NATURE_INJURY INJ_BODY_PART
## 1 CONTUSN,BRUISE,INTAC SKIN MULTIPLE PARTS (MORE THAN ONE MAJOR)
## 2 CUT,LACER,PUNCT-OPN WOUND HEAD,NEC
## 3 SPRAIN,STRAIN RUPT DISC ANKLE
## 4 SPRAIN,STRAIN RUPT DISC BACK (MUSCLES/SPINE/S-CORD/TAILBONE)
## 5 BURN,CHEMICL-FUME,COMPOUN HAND (NOT WRIST OR FINGERS)
## 6 SPRAIN,STRAIN RUPT DISC BACK (MUSCLES/SPINE/S-CORD/TAILBONE)
## SCHEDULE_CHARGE DAYS_RESTRICT DAYS_LOST TRANS_TERM RETURN_TO_WORK_DT
## 1 0 8 0 N 03/26/2012
## 2 0 0 0 N 1/09/2007
## 3 0 0 9 N 07/14/2009
## 4 NA 5 NA N 6/01/2000
## 5 0 5 0 N 2/01/2005
## 6 0 3 0 N 4/04/2006
## IMMED_NOTIFY INVEST_BEGIN_DT COAL_METAL_IND
## 1 NO VALUE FOUND M
## 2 NO VALUE FOUND M
## 3 NO VALUE FOUND M
## 4 NOT MARKED M
## 5 NO VALUE FOUND 01/22/2005 M
## 6 NO VALUE FOUND Mcolnames(df)## [1] "CONTROLLER_NAME" "OPERATOR_NAME" "CONTRACTOR_ID"
## [4] "SUBUNIT" "CAL_YR" "CAL_QTR"
## [7] "FISCAL_YR" "FISCAL_QTR" "ACCIDENT_TIME"
## [10] "DEGREE_INJURY" "FIPS_STATE_CD" "UG_LOCATION"
## [13] "UG_MINING_METHOD" "MINING_EQUIP" "EQUIP_MFR_NAME"
## [16] "EQUIP_MODEL_NO" "SHIFT_BEGIN_TIME" "CLASSIFICATION"
## [19] "ACCIDENT_TYPE" "NO_INJURIES" "TOT_EXPER"
## [22] "MINE_EXPER" "JOB_EXPER" "OCCUPATION"
## [25] "ACTIVITY" "INJURY_SOURCE" "NATURE_INJURY"
## [28] "INJ_BODY_PART" "SCHEDULE_CHARGE" "DAYS_RESTRICT"
## [31] "DAYS_LOST" "TRANS_TERM" "RETURN_TO_WORK_DT"
## [34] "IMMED_NOTIFY" "INVEST_BEGIN_DT" "COAL_METAL_IND"#apply(is.na(df), 2, sum) including the non missing valuesWe need to convert the missing values to NA
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionmutated_data <-
mutate(df,
CONTRACTOR_ID = na_if(CONTRACTOR_ID, ""),
ACCIDENT_TIME = ifelse(ACCIDENT_TIME >2400, NA, ACCIDENT_TIME),
UG_LOCATION = ifelse(UG_LOCATION =="NO VALUE FOUND", NA, UG_LOCATION),
UG_MINING_METHOD = ifelse(UG_MINING_METHOD =="NO VALUE FOUND", NA, UG_MINING_METHOD),
MINING_EQUIP = ifelse(MINING_EQUIP =="NO VALUE FOUND", NA, MINING_EQUIP),
EQUIP_MFR_NAME = ifelse(EQUIP_MFR_NAME =="NO VALUE FOUND", NA, EQUIP_MFR_NAME),
EQUIP_MODEL_NO = ifelse(EQUIP_MODEL_NO =="?" || EQUIP_MODEL_NO =="", NA, EQUIP_MODEL_NO),
SHIFT_BEGIN_TIME = ifelse(SHIFT_BEGIN_TIME >2400, NA, SHIFT_BEGIN_TIME),
TOT_EXPER = ifelse(TOT_EXPER =="", NA, TOT_EXPER),
MINE_EXPER = ifelse(MINE_EXPER =="", NA, MINE_EXPER),
JOB_EXPER = ifelse(JOB_EXPER =="", NA, JOB_EXPER),
OCCUPATION = ifelse(OCCUPATION =="NO VALUE FOUND", NA, OCCUPATION),
ACTIVITY = ifelse(ACTIVITY =="NO VALUE FOUND", NA, ACTIVITY),
INJURY_SOURCE = ifelse(INJURY_SOURCE =="NO VALUE FOUND", NA, INJURY_SOURCE),
NATURE_INJURY = ifelse(NATURE_INJURY =="NO VALUE FOUND", NA, NATURE_INJURY),
INJ_BODY_PART = ifelse(INJ_BODY_PART =="NO VALUE FOUND", NA, INJ_BODY_PART),
SCHEDULE_CHARGE = ifelse(SCHEDULE_CHARGE =="", NA, SCHEDULE_CHARGE),
DAYS_RESTRICT = ifelse(DAYS_RESTRICT =="", NA, DAYS_RESTRICT),
DAYS_LOST = ifelse(DAYS_LOST =="", NA, DAYS_LOST),
TRANS_TERM = ifelse(TRANS_TERM =="", NA, TRANS_TERM),
RETURN_TO_WORK_DT = ifelse(RETURN_TO_WORK_DT =="", NA, RETURN_TO_WORK_DT),
IMMED_NOTIFY = ifelse(IMMED_NOTIFY =="NO VALUE FOUND", NA, IMMED_NOTIFY),
INVEST_BEGIN_DT = ifelse(INVEST_BEGIN_DT =="", NA, INVEST_BEGIN_DT),
COAL_METAL_IND = ifelse(COAL_METAL_IND =="", NA, COAL_METAL_IND)
)## Warning in EQUIP_MODEL_NO == "?" || EQUIP_MODEL_NO == "": 'length(x) = 2000 > 1'
## in coercion to 'logical(1)'
## Warning in EQUIP_MODEL_NO == "?" || EQUIP_MODEL_NO == "": 'length(x) = 2000 > 1'
## in coercion to 'logical(1)'Checking the number of null values now after converting “?” and “VALUE NOT FOUND” to NA
count_missing <- function(df) {
sapply(df, FUN = function(col) sum(is.na(col)) )
}
nacounts <- count_missing(mutated_data)
hasNA = which(nacounts > 0)
nacounts[hasNA]## CONTRACTOR_ID ACCIDENT_TIME UG_LOCATION UG_MINING_METHOD
## 1830 139 1284 1392
## MINING_EQUIP EQUIP_MFR_NAME SHIFT_BEGIN_TIME TOT_EXPER
## 1053 1056 94 360
## MINE_EXPER JOB_EXPER OCCUPATION ACTIVITY
## 333 328 242 239
## INJURY_SOURCE NATURE_INJURY INJ_BODY_PART SCHEDULE_CHARGE
## 238 242 242 673
## DAYS_RESTRICT DAYS_LOST TRANS_TERM RETURN_TO_WORK_DT
## 543 422 267 322
## IMMED_NOTIFY INVEST_BEGIN_DT
## 1139 1530By looking at the dataset after replacing all the sentinel values with NA we see that CONTRACTOR_ID column has most of its values as NA so it is safe to drop it
colnames(mutated_data)## [1] "CONTROLLER_NAME" "OPERATOR_NAME" "CONTRACTOR_ID"
## [4] "SUBUNIT" "CAL_YR" "CAL_QTR"
## [7] "FISCAL_YR" "FISCAL_QTR" "ACCIDENT_TIME"
## [10] "DEGREE_INJURY" "FIPS_STATE_CD" "UG_LOCATION"
## [13] "UG_MINING_METHOD" "MINING_EQUIP" "EQUIP_MFR_NAME"
## [16] "EQUIP_MODEL_NO" "SHIFT_BEGIN_TIME" "CLASSIFICATION"
## [19] "ACCIDENT_TYPE" "NO_INJURIES" "TOT_EXPER"
## [22] "MINE_EXPER" "JOB_EXPER" "OCCUPATION"
## [25] "ACTIVITY" "INJURY_SOURCE" "NATURE_INJURY"
## [28] "INJ_BODY_PART" "SCHEDULE_CHARGE" "DAYS_RESTRICT"
## [31] "DAYS_LOST" "TRANS_TERM" "RETURN_TO_WORK_DT"
## [34] "IMMED_NOTIFY" "INVEST_BEGIN_DT" "COAL_METAL_IND"DropColumns<- mutated_data
drop <- c("CONTRACTOR_ID")
mutated_data = DropColumns[,!(names(DropColumns) %in% drop)]
colnames(mutated_data)## [1] "CONTROLLER_NAME" "OPERATOR_NAME" "SUBUNIT"
## [4] "CAL_YR" "CAL_QTR" "FISCAL_YR"
## [7] "FISCAL_QTR" "ACCIDENT_TIME" "DEGREE_INJURY"
## [10] "FIPS_STATE_CD" "UG_LOCATION" "UG_MINING_METHOD"
## [13] "MINING_EQUIP" "EQUIP_MFR_NAME" "EQUIP_MODEL_NO"
## [16] "SHIFT_BEGIN_TIME" "CLASSIFICATION" "ACCIDENT_TYPE"
## [19] "NO_INJURIES" "TOT_EXPER" "MINE_EXPER"
## [22] "JOB_EXPER" "OCCUPATION" "ACTIVITY"
## [25] "INJURY_SOURCE" "NATURE_INJURY" "INJ_BODY_PART"
## [28] "SCHEDULE_CHARGE" "DAYS_RESTRICT" "DAYS_LOST"
## [31] "TRANS_TERM" "RETURN_TO_WORK_DT" "IMMED_NOTIFY"
## [34] "INVEST_BEGIN_DT" "COAL_METAL_IND"nacounts <- count_missing(mutated_data)
hasNA = which(nacounts > 0)
nacounts[hasNA]## ACCIDENT_TIME UG_LOCATION UG_MINING_METHOD MINING_EQUIP
## 139 1284 1392 1053
## EQUIP_MFR_NAME SHIFT_BEGIN_TIME TOT_EXPER MINE_EXPER
## 1056 94 360 333
## JOB_EXPER OCCUPATION ACTIVITY INJURY_SOURCE
## 328 242 239 238
## NATURE_INJURY INJ_BODY_PART SCHEDULE_CHARGE DAYS_RESTRICT
## 242 242 673 543
## DAYS_LOST TRANS_TERM RETURN_TO_WORK_DT IMMED_NOTIFY
## 422 267 322 1139
## INVEST_BEGIN_DT
## 1530dimension <- c(ncol(mutated_data),nrow(mutated_data))
dimension## [1] 35 2000We see that column SHIFT_BEGIN_TIME has 94 null rows so it is safe to drop the rows
newdata <-
mutated_data[!is.na(mutated_data$SHIFT_BEGIN_TIME),]
nrow(mutated_data)## [1] 2000nrow(newdata)## [1] 1906nacounts <- count_missing(newdata)
hasNA = which(nacounts > 0)
nacounts[hasNA]## ACCIDENT_TIME UG_LOCATION UG_MINING_METHOD MINING_EQUIP
## 91 1221 1319 984
## EQUIP_MFR_NAME TOT_EXPER MINE_EXPER JOB_EXPER
## 987 311 285 282
## OCCUPATION ACTIVITY INJURY_SOURCE NATURE_INJURY
## 211 209 209 210
## INJ_BODY_PART SCHEDULE_CHARGE DAYS_RESTRICT DAYS_LOST
## 210 619 499 386
## TRANS_TERM RETURN_TO_WORK_DT IMMED_NOTIFY INVEST_BEGIN_DT
## 237 284 1111 1464newdata2 <- newdata[!is.na(newdata$ACCIDENT_TIME),]
nrow(mutated_data)## [1] 2000nrow(newdata2)## [1] 1815after removing the rows from ACCIDENT_TIME which were 87 rows and it was safe to remove them
dimension <- c(ncol(newdata2),nrow(newdata2))
dimension## [1] 35 1815nacounts <- count_missing(newdata2)
hasNA = which(nacounts > 0)
nacounts[hasNA]## UG_LOCATION UG_MINING_METHOD MINING_EQUIP EQUIP_MFR_NAME
## 1169 1265 921 923
## TOT_EXPER MINE_EXPER JOB_EXPER OCCUPATION
## 276 250 249 182
## ACTIVITY INJURY_SOURCE NATURE_INJURY INJ_BODY_PART
## 180 180 181 181
## SCHEDULE_CHARGE DAYS_RESTRICT DAYS_LOST TRANS_TERM
## 601 483 376 203
## RETURN_TO_WORK_DT IMMED_NOTIFY INVEST_BEGIN_DT
## 249 1056 1396Now we replace the variables that numeric with the mean of the column
library(dplyr)
new_data3 <-
mutate(newdata2,
TOT_EXPER = ifelse(is.na(TOT_EXPER),format(round(mean(newdata2$TOT_EXPER, na.rm = TRUE), 3), nsmall = 3) , TOT_EXPER),
MINE_EXPER = ifelse(is.na(MINE_EXPER), format(round(mean(newdata2$MINE_EXPER, na.rm = TRUE), 3), nsmall = 3), MINE_EXPER),
JOB_EXPER = ifelse(is.na(JOB_EXPER), format(round(mean(newdata2$JOB_EXPER, na.rm = TRUE), 3), nsmall = 3), JOB_EXPER), )
nacounts <- count_missing(new_data3)
hasNA = which(nacounts > 0)
nacounts[hasNA]## UG_LOCATION UG_MINING_METHOD MINING_EQUIP EQUIP_MFR_NAME
## 1169 1265 921 923
## OCCUPATION ACTIVITY INJURY_SOURCE NATURE_INJURY
## 182 180 180 181
## INJ_BODY_PART SCHEDULE_CHARGE DAYS_RESTRICT DAYS_LOST
## 181 601 483 376
## TRANS_TERM RETURN_TO_WORK_DT IMMED_NOTIFY INVEST_BEGIN_DT
## 203 249 1056 1396Now we process the Categorical values and replace the NA values with a variable which makes sense
library(dplyr)
new_data4 <-
mutate(new_data3,
UG_LOCATION = ifelse(is.na(UG_LOCATION), "UNKNOWN_UG_LOCATION" , UG_LOCATION),
UG_MINING_METHOD = ifelse(is.na(UG_MINING_METHOD), "UNKNOWN_UG_MINING_METHOD" , UG_MINING_METHOD),
MINING_EQUIP = ifelse(is.na(MINING_EQUIP), "INVALID_EQUIPMENT" , MINING_EQUIP),
EQUIP_MFR_NAME = ifelse(is.na(EQUIP_MFR_NAME), "INVALID_MFR_CODE", EQUIP_MFR_NAME),
OCCUPATION = ifelse(is.na(OCCUPATION), "INVALID_OCCUPATION", OCCUPATION),
ACTIVITY = ifelse(is.na(ACTIVITY), "INVALID_ACTIVITY", ACTIVITY),
INJURY_SOURCE = ifelse(is.na(INJURY_SOURCE), "INVALID_INJURY", INJURY_SOURCE),
NATURE_INJURY = ifelse(is.na(NATURE_INJURY), "INVALID_NATURE_INJURY", NATURE_INJURY),
INJ_BODY_PART = ifelse(is.na(INJ_BODY_PART), "INVALID_INJ_BODY_PART", INJ_BODY_PART),
SCHEDULE_CHARGE = ifelse(is.na(SCHEDULE_CHARGE), format(round(mean(new_data3$SCHEDULE_CHARGE, na.rm = TRUE), 3), nsmall = 3), SCHEDULE_CHARGE),
DAYS_RESTRICT = ifelse(is.na(DAYS_RESTRICT), 0, DAYS_RESTRICT),
# Since in the data dictionary or the details of the variables it is clearly stated
# that if there are null values in this case then that means the DAYS_RESTRICT would be 0 for the null values which do not apply
DAYS_LOST = ifelse(is.na(DAYS_LOST), 0, DAYS_LOST),
# Since in the data dictionary or the details of the variables it is clearly stated
# that if there are null values in this case then that means the DAYS_LOST would be 0 for the null values which do not apply
TRANS_TERM = ifelse(is.na(TRANS_TERM), "INVALID_TRANS_TERM", TRANS_TERM),
RETURN_TO_WORK_DT = ifelse(is.na(RETURN_TO_WORK_DT), "INVALID_RETURN_TO_WORK_DT", RETURN_TO_WORK_DT),
IMMED_NOTIFY = ifelse(is.na(IMMED_NOTIFY), "INVALID_NOTIFICATION", IMMED_NOTIFY),
INVEST_BEGIN_DT = ifelse(is.na(INVEST_BEGIN_DT), "INVALID_INVEST_BEGIN_DT", INVEST_BEGIN_DT),
)
#CONVERTING THE RETURN_TO_WORK_DT from date format to year format
nacounts <- count_missing(new_data4)
hasNA = which(nacounts > 0)
nacounts[hasNA]## named integer(0)This point is at the ending of the DATA CLEANING AND PROCESSING stage now we visualise with the variables we are left with after Cleaning and Processing the data
Exploratory Data Analysis with one variable
Bar charts
Firstly we try to find the frequency of the DEGREE_INJURY column to check the severity of the injury of the accident using a Sorted Horizontal Bar Charts
library(ggplot2)
#Over here we see that there needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
statesums <- table(new_data4$DEGREE_INJURY)
statef <- as.data.frame(statesums)
colnames(statef) <- c("DEGREE_INJURY", "count")
statef <- transform(statef,
DEGREE_INJURY=reorder(DEGREE_INJURY, -desc(count)))
g <- ggplot(statef,
aes(x = DEGREE_INJURY, y = count, label = count, fill="maroon"))+ coord_flip() + geom_col() + geom_text(nudge_y = -1, color = "black") + ggtitle("DEGREE_INJURY against count") + xlab("DEGREE_INJURY") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(hjust = 0.8))
g
In the Plot above we get to see that the severity of the injuries or the degree of injuries where the most of all the categories the employees fell in the category where they had Days away from work only, “No days away from work and no restricted activities” and “Days restricted activity only” which means employees mostly were away from work and some even did not get days off and action restrictions in the work area and some of them had days restricted activity only. Other categories include “Accidents only” and “Days away from work and restricted actions” which means they faced just the accidents and days off from work with limited activities in the work environment. Other minor categories included “Fatality”,“Occupational Illness”.
Next we will explore the UG_LOCATION column to check where the accident occured or where the employee was when the accident occured
library(ggplot2)
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
statesums <- table(new_data4$UG_LOCATION)
statef <- as.data.frame(statesums)
colnames(statef) <- c("UG_LOCATION", "count")
statef <- transform(statef,
UG_LOCATION=reorder(UG_LOCATION, -desc(count)))
g <- ggplot(statef,
aes(x = UG_LOCATION, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("UG_LOCATION against count") +
xlab("UG_LOCATION") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(hjust = 0.8))
g
In the plot above we get get to see there is a suspiciously high count for UNKNOWN_UG_LOCATION meaning the underground location is unknown to us. But if we see the corresponding SUBUNIT column we can deduce something so we plot the barchart for the subunits to see which subunit the accident occured
library(ggplot2)
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
statesums <- table(new_data4$SUBUNIT)
statef <- as.data.frame(statesums)
colnames(statef) <- c("SUBUNIT", "count")
statef <- transform(statef,
SUBUNIT=reorder(SUBUNIT, -desc(count)))
g <- ggplot(statef,
aes(x = SUBUNIT, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("SUBUNIT against count") +
xlab("SUBUNIT") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(hjust = 0.8))
g
Although we see that the location might be in the underground but after seeing the subunit we can deduce that although the location might be unknown but for subunits most of the accident occured were mainly in the Strip,Quarry,Open Pit And Mill Operation/Preparation Plant and in the Underground subunit.
Next we can explore the UG_MINING_METHOD or the Underground mining method
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
statesums <- table(new_data4$UG_MINING_METHOD)
statef <- as.data.frame(statesums)
colnames(statef) <- c("UG_MINING_METHOD", "count")
statef <- transform(statef,
UG_MINING_METHOD=reorder(UG_MINING_METHOD, -desc(count)))
g <- ggplot(statef,
aes(x = UG_MINING_METHOD, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("UG_MINING_METHOD against count") +
xlab("UG_MINING_METHOD") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(hjust = 0.8))
g
We have seem to come across anomalies where it shows that the Mining Methods are missing for the maximum counts so the mining methods are unknown for most of the cases maybe if we explore other columns we maybe able to deduce what this column is trying to tell us
Now we see the MINING_EQUIP column to see what Mining Equipments were there in the accident
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
df <- as.data.frame(table(new_data4$MINING_EQUIP))
df <- df[order(-df$Freq),]
dfshort <- df[1:20,]
statesums <- dfshort
statef <- statesums
colnames(statef) <- c("MINING_EQUIP", "count")
statef <- transform(statef,
MINING_EQUIP=reorder(MINING_EQUIP, -desc(count)))
g <- ggplot(statef,
aes(x = MINING_EQUIP, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("MINING_EQUIP against count") +
xlab("MINING_EQUIP") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(hjust = 0.7))
g
Since the number of MINING_EQUIPs were huge we had to cut it down to top
20 values in order to plot a clear bar chart Unfortunately there are so
many categories that horizontal barcharts might not be appropriate but
we can still see that most of the equipments during the accident were
Invalid Equipments as it was mentioned that any missing values in the
column would be considered as an invalid equipment according to the data
dictionary.
So up until this stage we found out that most of the accidents occured in unknow Underground locations with unknown mining methods with invalid equipments.
Next we explore the CLASSIFICATION column
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
df <- as.data.frame(table(new_data4$OCCUPATION))
df <- df[order(-df$Freq),]
dfshort <- df[1:20,]
statesums <- dfshort
statef <- statesums
colnames(statef) <- c("CLASSIFICATION", "count")
statef <- transform(statef,
CLASSIFICATION=reorder(CLASSIFICATION, -desc(count)))
g <- ggplot(statef,
aes(x = CLASSIFICATION, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("CLASSIFICATION against count") +
xlab("CLASSIFICATION") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(size = rel(0.6) ))+
theme(axis.text.y=element_text(size=rel(0.6)))
g
Since the number of CLASSIFICATION was huge we had to reduce it down to
first 20 values in order to perform decent plotting We can deduce that
most of the Classification falls on Handling of materials, Slip or fall
of person, Non powered handtools, machinery, fall of roof or back and
power haulage. So most of the Accidents happened due to Handling of
materials
We can be more clear if we explore the ACCIDENT_TYPE column
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
statesums <- table(new_data4$ACCIDENT_TYPE)
statef <- as.data.frame(statesums)
colnames(statef) <- c("ACCIDENT_TYPE", "count")
statef <- transform(statef,
ACCIDENT_TYPE=reorder(ACCIDENT_TYPE, -desc(count)))
g <- ggplot(statef,
aes(x = ACCIDENT_TYPE, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("ACCIDENT_TYPE against count") +
xlab("ACCIDENT_TYPE") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(size = rel(0.7) ))+
theme(axis.text.y=element_text(size=rel(0.6)))
g
In the plot above we see that most of the Accident types from the ACCIDENT_TYPE columns is the N.E.C classified Accidents meaning when an accident is classified as bodily reaction and exertion, n.e.c., it refers to an accident that resulted from a combination of free bodily motion and excessive physical effort or action. It is a type of bodily reaction and exertion accident. However, it is a bodily reaction and exertion accident that does not fit or fall into a category. It is classified as n.e.c.. This means it is not elsewhere classified. Other accident types include the ones without injuries and over exertion in lifting objects. Other down the row we see accidents such as getting struck by a stationary objects and caught in, under and below moving objects.
Which brings a question whether the employees who faced the most of the accidents were inexperienced or experienced
So now we explore the column which gives us the information about the employee’s occupation from the OCCUPATION column
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
df <- as.data.frame(table(new_data4$OCCUPATION))
df <- df[order(-df$Freq),]
dfshort <- df[1:20,]
statesums <- dfshort
statef <- statesums
colnames(statef) <- c("OCCUPATION", "count")
statef <- transform(statef,
OCCUPATION=reorder(OCCUPATION, -desc(count)))
g <- ggplot(statef,
aes(x = OCCUPATION, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("OCCUPATION against count") +
xlab("OCCUPATION") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(size = rel(0.7) ))+
theme(axis.text.y=element_text(size=rel(0.6)))
g
Since the number of OCCUPATIONs were huge we had to cut it down to first
20 values in order to do comparison In this plot above we see that most
of the occupation from the Bar chart of the OCCUPATION are from the
group Maintanence man, Mechanic, Serviceman/Repair,Boilermaker,Fueler,
Tire tech, Field Service tech and the second most occuring group is an
Invalid group which raises the question, What Occupation were they
invovled with? Other columns included Warehouse Bagger, Store Keeper,
Packager, Fabricator, Cleaning Plant Operator
Next we explore the activities that they were doing at that time so we check the ACTIVITIES column
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
df <- as.data.frame(table(new_data4$ACTIVITY))
df <- df[order(-df$Freq),]
dfshort <- df[1:20,]
statesums <- dfshort
statef <- statesums
colnames(statef) <- c("ACTIVITY", "count")
statef <- transform(statef,
ACTIVITY=reorder(ACTIVITY, -desc(count)))
g <- ggplot(statef,
aes(x = ACTIVITY, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("ACTIVITY against count") +
xlab("ACTIVITY") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(size = rel(0.7) ))+
theme(axis.text.y=element_text(size=rel(0.6)))
g
Since the number of ACTIVITY was very huge we had to cut it down to
first 20 values in order to do comparison We find out from the above
plot that most of the activities were Handling Supplies Materials,
Machine Maintanence Repairing,Handling Non Powered Tools, Invalid
Activities this raises another question whether they were experienced or
not ao to further clear our doubts we look at INJURY_SOURCE column
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
df <- as.data.frame(table(new_data4$INJURY_SOURCE))
df <- df[order(-df$Freq),]
dfshort <- df[1:10,]
statesums <- dfshort
statef <- statesums
colnames(statef) <- c("INJURY_SOURCE", "count")
statef <- transform(statef,
INJURY_SOURCE=reorder(INJURY_SOURCE, -desc(count)))
g <- ggplot(statef,
aes(x = INJURY_SOURCE, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("INJURY_SOURCE against count") +
xlab("INJURY_SOURCE") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(size = rel(0.7) ))+
theme(axis.text.y=element_text(size=rel(0.6)))
g
Since the amount of INJURY_SOURCE is so huge that we took the top 10 of
the values and plotted our graph
In the above plot most of the injury happened from metal pipe wire nails, invalid injuries, metal covers and guards, the ground broken rock coal and caving rock coal. Further Single Column investigation is needed to Find what the invalid columns are and also to find a link to Injuries or accidents
#Over here we see that there also needs a reordering of the values in order to find a clearer output
#So we sort the values alphabetically in the increasing order and added default colors to pick all the variables
statesums <- table(new_data4$NATURE_INJURY)
statef <- as.data.frame(statesums)
colnames(statef) <- c("NATURE_INJURY", "count")
statef <- transform(statef,
NATURE_INJURY=reorder(NATURE_INJURY, -desc(count)))
g <- ggplot(statef,
aes(x = NATURE_INJURY, y = count, label = count, fill="maroon"))+ coord_flip() +
geom_col() +
geom_text(nudge_y = -1, color = "black") +
ggtitle("NATURE_INJURY against count") +
xlab("NATURE_INJURY") + ylab("count") +
theme_bw() + theme(legend.position = "none") +
theme(plot.title = element_text(size = rel(0.7) ))+
theme(axis.text.y=element_text(size=rel(0.6)))
g
in the above plot most of the nature of the injury falls under Open Wounds, Fractured Chip and Bruises and some even have Bruises and there are some that fall into Invalid Injury sources. So it is safe to consider the invalid injury sources has missing since they might not be working there in the mining sectors.
Histogram
We would look into the next column of interest which is DAYS_LOST so for this we will be using a histogram
ggplot(new_data4, aes(x=DAYS_LOST)) +
geom_histogram(binwidth=10, fill="Gray")+ggtitle("DAYS_LOST against count")+theme(plot.title = element_text(size = rel(0.7) ))
#BINS need to be wider to make it less noisyWe can see in the plot above that there are outliers at 200 to 360 days and one at 450 and one at 500 and most of the concentration is at around 7 days and it is suspiciously very high around this are but a histogram of this plot is not much clear as to what meaning we should find out from this so we can use a Density Plot for this column
library(scales)
##Since this the above code gives a scewed graph we therefore use logarithms
#density plot was positively skewed so I used logarithms
# Basic density plot in ggplot2
ggplot(new_data4, aes(x = DAYS_LOST)) +
geom_density() + scale_x_log10(breaks=c(2.5,5,45,360,450,1000),labels=comma) +
annotation_logticks(sides="bt") + theme(text = element_text(size = 10))+ggtitle("DAYS_LOST against count")+theme(plot.title = element_text(size = rel(0.7) ))## Warning: Transformation introduced infinite values in continuous x-axis## Warning: Removed 1119 rows containing non-finite values (stat_density).
We see that there are two peaks where one is at 2.5 days and the other one is at 45 days lost so these are the maximum days that were lost as a consequence to the accident
Exploratory Data Analysis with two variables
Now it is time for the comparison of two or more variables We will look at the columns UG_LOCATION against DEGREE_INJURY to see a relation
Heat maps
ggplot(data = new_data4) +ggtitle("CAL_YR against DEGREE_INJURY")+theme(plot.title = element_text(size = rel(0.7) )) +
geom_count(mapping = aes(x=CAL_YR, y=DEGREE_INJURY))+theme(axis.text.x = element_text(size = 5))+theme(axis.text.x = element_text(angle = 30, vjust = 1, hjust = 1)) 
We can clearly state that the DEGREE_INJURY decreases accross the accident years which is CAL_YR since density decreases across the years
Next we take a look at the relationship between DEGREE_INJURY and JOB_EXPER through a scatter plot
#degree injury against job experience
ggplot(data= new_data4, mapping = aes(JOB_EXPER,DEGREE_INJURY)) + ggtitle("JOB_EXPER against DEGREE_INJURY")+theme(plot.title = element_text(size = rel(0.7) )) + geom_point(aes(as.numeric(JOB_EXPER),DEGREE_INJURY,color= DEGREE_INJURY), show.legend = FALSE)
We can clearly state that as JOB_EXPER or job experience increases the injury degree or DEGREE_INJURY decreases
Density plots and Heat maps
require(gridExtra)## Loading required package: gridExtra##
## Attaching package: 'gridExtra'## The following object is masked from 'package:dplyr':
##
## combine#shift time against accident time
g1<-ggplot(data= new_data4, mapping = aes(SHIFT_BEGIN_TIME,ACCIDENT_TIME)) + ggtitle("SHIFT_BEGIN_TIME against ACCIDENT_TIME") +theme(plot.title = element_text(size = rel(0.7) ))+ geom_point(aes(SHIFT_BEGIN_TIME,ACCIDENT_TIME,color= SHIFT_BEGIN_TIME), show.legend = FALSE)
g2<-ggplot(data = new_data4) +
geom_count(mapping = aes(x=SHIFT_BEGIN_TIME, y=ACCIDENT_TIME))+theme(axis.text.x = element_text(size = 5))+theme(axis.text.x = element_text(angle = 90, vjust = 1, hjust = 1))+theme(axis.text.y = element_text(angle = 90, vjust = 1, hjust = 1))
grid.arrange(g1, g2, nrow=2)
The plot for both the graphs. The heat map and density plot shows that the accidents occured more between shifts of 6000 which is 6am and 7000 which is 7am
ggplot(data = new_data4) +
geom_count(mapping = aes(x=ACCIDENT_TIME, y=OCCUPATION))+theme(axis.text.x = element_text(size = 5))+theme(axis.text.y = element_text(size = 5))+theme(axis.text.y = element_text(angle = 0, vjust = 1, hjust = 1))+ggtitle("ACCIDENT_TIME against OCCUPATION") +theme(plot.title = element_text(size = rel(0.7) ))
Referring to the plot above the Occupation of Maintenance man,Mechanic,
Repair/Serviceman,Boilermaker etc has the most dense ACCIDENT_TIME
between approximately 7500 and 1500 of the 24 hour clock of the
ACCIDENT_TIME axis and we also get to see that the INVALID_OCCUPATION of
the OCCUPATION axis had uniform accident in the ACCIDENT_TIME axis
throughout from 0000 to 2400 with the same density of the heat map. We
can assume that the INVALID_OCCUPATION falls well underneath occupations
which include excessive physical work like a movie set where stunt
people practice their movie shots.